I-Have-ADHD List Item Count Limit: Why Responses Max Out at 5 Items

The i-have-adhd skill enforces a strict maximum of five items per list to keep responses concise and accessible for readers with ADHD.

The ayghri/i-have-adhd repository implements a specialized set of prompting rules designed to optimize AI outputs for cognitive accessibility. At the core of this system lies a non-negotiable constraint on list length that directly shapes how information is structured and delivered.

The 5-Item Hard Cap Rule

The limitation is explicitly codified in skills/i-have-adhd/SKILL.md at lines 103-106, under the directive "Cap lists at 5 items". This is not a suggestion—it is a mandatory architectural boundary that all generated content must respect.

According to the source code, this constraint serves a specific cognitive purpose. Long, unbroken lists create cognitive overload for readers with ADHD, making it difficult to prioritize, retain, or act on presented information. The five-item threshold represents a calculated balance between completeness and digestibility.

How the Limit Is Enforced in Practice

When a response naturally requires more than five points, the skill mandates list splitting rather than truncation. This preserves all information while respecting the structural boundary. Common partitioning strategies include:

• Temporal splits: "Do now" versus "Do later"
• Priority splits: "Must have" versus "Nice to have"
• Complexity splits: "Basic steps" versus "Advanced options"

Example: Single List Within Limit

When five or fewer items suffice, the list renders unchanged:

1. Open `src/auth.ts`
2. Replace `verifyToken` (lines 42-58) with the new snippet
3. Run `npm test -- auth.spec.ts`
4. Verify that the login flow works locally
5. Commit the changes

Example: Split List Exceeding Limit

When the content demands more than five items, the model must restructure:

**Do now (first 5 items)**
1. Open `src/auth.ts`
2. Replace `verifyToken` (lines 42-58)
3. Update the token validation logic
4. Run `npm test -- auth.spec.ts`
5. Verify that the login flow works locally

**Do later**
1. Add unit tests for edge-case tokens
2. Document the new auth flow in `README.md`
3. Update CI pipeline to include auth tests

Where the Rule Is Defined

File	Purpose
skills/i-have-adhd/SKILL.md	Canonical skill definition containing the "Cap lists at 5 items" rule at lines 103-106
.cursor/skills/i-have-adhd/SKILL.md	Cursor IDE-compatible mirror of the canonical skill file

Both files remain synchronized, ensuring consistent behavior across deployment contexts. The rule appears in no other configuration files—README.md and AGENTS.md describe system behavior without restating specific numeric constraints.

Summary

• Hard limit: Five items maximum per list, as defined in SKILL.md
• Enforcement mechanism: Mandatory list splitting when content exceeds the threshold
• Rationale: Prevents cognitive overload and supports ADHD accessibility needs
• Implementation: Partition strategies like "now/later" or "must/nice to have"

Frequently Asked Questions

What happens if a response needs more than five items?

The skill requires the model to split the list into multiple categorized sections rather than presenting a single long list. Each resulting section must independently respect the five-item cap.

Is the five-item limit configurable?

No. The limit is hardcoded in the canonical skill file at skills/i-have-adhd/SKILL.md and appears to be a fixed architectural constraint rather than a user-tunable parameter.

Does this limit apply to numbered and bulleted lists equally?

Yes. The "Cap lists at 5 items" rule applies to all list formats regardless of marker style. The constraint concerns item count, not presentation format.

Why five items specifically?

The repository does not explicitly document the psychological research behind this threshold, but the implementation suggests it represents an empirically-derived boundary for working memory capacity in readers with ADHD—balancing information density with cognitive load management.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}